{
  "id": 4093,
  "url": "https://arxiv.org/abs/2605.18714v2",
  "title": "Semantic Generative Tuning for Unified Multimodal Models",
  "summary": "Unified multimodal models (UMMs) strive to consolidate visual understanding and visual generation within a single architecture. However, prevailing training paradigms independently optimize understanding via sparse text signals and generation through dense pixel objectives. Such a decoupled strategy yields misaligned representation spaces, isolating visual understanding from generation and hindering their mutual reinforcement. This work presents the first systematic investigation into generative",
  "authors": "Songsong Yu, Yuxin Chen, Ying Shan, Yanwei Li",
  "category": "research",
  "topics": "safety-alignment,finance-investment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-18T17:46:46.000Z",
  "fetched_at": "2026-07-14T16:30:45.938Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/4093",
  "original_url": "https://arxiv.org/abs/2605.18714v2",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}